Transit is programmable, versioned data infrastructure for applications. You
declare data (Tables, Key-Value maps, collaborative Documents, and external
datasets), computations (point-in-time queries, live subscriptions,
maintained derived views, and stateful dataflows), and policies (scoped
authorization, retention, capacity, and placement). Transit's runtime activates
that data through authorized publication, query execution and durable processes
(Station); connectors coordinate recorded external effects. The installed
local runtime persists native history in write-ahead logs and immutable state in
local objects. Its direct Table owner accepts data independently of actor turns,
retains unflushed data for queries and publishes .tseg objects for bulk writes.
Automatic point-write flushing, tiered caching and automatic incremental
activation remain extensions to that installed path.
Start with the data your application needs, what you want to compute from it, and the policies governing its use. The data model connects those declarations to their public interfaces and observable outcomes, and Internals formalizes the underlying system, storage, query, state, and execution models.
Read the published handbook, starting with What is Transit?, the data model and your first dataset.
| Reference | Entry point |
|---|---|
| Data and policies | Data model, security, retention |
| Query language | SQL |
| Commands and recovery | Transit CLI |
| SQL connections | Server-managed sessions, PostgreSQL and MySQL |
| HTTP APIs | Serving OpenAPI, Board OpenAPI |
| Reactive streams | AsyncAPI |
| Analytical clients | Flight SQL and ADBC |
| Application clients | Language SDKs, native protocol |
| Workload design | Architectural priorities, performance and improvement plan, declarative benchmarks |
| System internals | Architecture, rationale and evidence, scientific standard |
| Interface coverage | Usage, contracts and internals map, schema downloads |
| Example recipes | SQL, Python, TypeScript, Rust and Zig |
The documentation index explains navigation and source ownership; the site guide covers local previews.
Use the contributor guide for source ownership and
validation responsibilities, the toolchain setup to prepare the
checkout, and the launcher guide for development lifecycle
commands. To run an installed bundle, follow the local quickstart
with transit on PATH.
Copy transit.yaml.example to transit.yaml for an annotated
local configuration covering services, SQL gateways, accounts, TLS and limits.
Every setting is commented out; uncomment only the keys you want to override,
including their parent sections. The repository ignores transit.yaml;
PostgreSQL and MySQL remain disabled until selected.
Follow the configuration reference
and SQL access guide when enabling them.
- Component map: libraries, apps, and their READMEs.
- Examples: complete applications and public SDK callers.
- Proof coverage: checks, failure schedules, and qualification limits.
- Agent instructions: how to change this repository.
For downstream application development, the local installation guide covers separately built Zig workers, scoped caller connections and compatible replacement through a language-neutral bootstrap contract.